ELSEVIER Energy and Buildings 24 ( 1996) l-10

The role of thermal mass on the cooling load of buildings. An overview

of computational methods
C.A. Balaras
N&mu1 Obsenutory oj’Athens. Institute ofMeteorology and Physics of the Atmospheric Environment, Group Energ)) Conserwtion. PO Box 20048,
GR I18 lOAthem, Greece

Accepted 19 June 1995

Abstract

Thermal mass can reduce peak cooling loads and indoor air temperature swings in buildings. The factors affecting the performance of
thermal mass are reviewed and classified. Experimental studies which demonstrate the effectiveness of thermal mass as an energy conservation
alternative and in providing more comfortable indoor conditions, are also reviewed. A number of simplified design tools for calculating the
cooling load and indoor air temperatures of a building, which also account for thermal mass effects, have been identified and are classified.
The models are categorized in terms of their inputs, outputs and restrictions on their level of accuracy in treating thermal mass effects, type
of loads or other design limitations.

Keywrdst Thermal mass; Coolmg load; Computational methods

1. Introduction
Indoor air temperatureis primarily influenced by external
climatological parameters(solar radiation, outdoor temper-
ature) and highly variable internal loads (human activity,
lights, equipment). During the summer,this resultsin tem-
perature swings,with peaksoccurring around noon andearly
afternoon hours. During theseperiods of the day, high out-
door temperatureseliminate the use of simple alternative
passive cooling techniques like natural ventilation. As a
result, in order to maintain thermal comfort during these
hours, it is necessaryto remove all excessheat from the space
immediately upon entry, by anenergy consumingmechanical
system. This will require an oversized system capable of
handling the short period peaksof the cooling load.
To reduce indoor air temperatureand cooling load peaks,
and to transfer the load to a later time in the day, it is possible
to store heat in the material of the outer envelope and the
interior mass of the building. The storage material is the
construction massof the building itself, which is referred to
as thermal mass.It is typically contained in walls, partitions,
ceilings and floors of the building, constructed of material
with high heat capacity, suchas poured concrete, bricks and
tiles.
The thermal massof the building could have a positive
effect on the indoor conditions during the summerand winter
periods.The energy available from the high solargainsduring
the day is stored and then is slowly releasedinto the indoor
environment at a later time. In winter, the stored heat is
transferred back into the room during the late afternoon and
late evening hours, when it is most needed,satisfying part of
the heating load and avoiding overheating and discomfort
conditions during the high solarradiation periodsof the day.
In summer,heatis storedin the thermal mass,thusreducing
the cooling load peaks. A reduced portion of the load will
needto be removed from the interior space,while the remain-
ing portion of the external and internal heatgainsis contained
within the thermal storagematerials.The stored heat is pro-
gressively releasedto the interior of the building. As a result
indoor air temperaturevariations remain within the comfort
range, for most hours during the day. This then reducesthe
energy consumption for cooling, which can reach 20% in
commercial buildings [ 11.
There is alsoa time shift of the peak load and a time lag of
the heat releasefrom the material to the indoor air. This time
lag is actually desirable,since at the time when the heat is
returned to the interior, the indoor temperature is relatively
lower. In the event that the building is alsounoccupiedduring
the evening hours, like office buildings, it is possibleto relax
the restrictions on indoor thermal comfort. In any event, out-
door conditions aremore favourable for passivecooling tech-
niques,like natural, hybrid or indirect ventilation, which can
be usedto remove portions of the load [ 21.

037%7788/96/$15.00 0 1996 Elsevier Science S.A. All rights reserved

SSDIO378-7788(95)00956-X
 The effectiveness of thermal storage is acceptable where
the diurnal variation of ambient temperatures exceeds 10 K
[ 31. Under such circumstances, ample night ventilation and
closing the building during the day, due to the thermal mass
effect, it is possible to achieve a temperature difference,
between the mean indoor and outdoor temperatures, greater
than 10 K in the negative direction and an indoor temperature
swing of 2.5 K.
Thermal mass has also a positive effect on occupant com-
fort. High mass buildings prohibit high interior air and wall
temperature variations and sustain a more steady overall ther-
mal environment. This increases comfort, particularly during
mild seasons (spring and fall), during large air temperature
changes (high solar gain), and in areas with large day-night
temperature swings [ 41. Traditional architecture, especially
in southern European countries, has successfully demon-
strated these positive effects.
2. Parameters affecting the performance of thermal
mass
The rate of heat transfer through and the effectiveness of
thermal massare determinedby a number of parametersand
conditions. To achieve the bestpossibleresults,one needsto
follow somegeneralguidelines and take appropriateactions
that fall within the overall procedure for energy efficient
building. It is important though to alsounderstandthe relation
of these parameterson the performance of thermal massin
order to achieve the best possible results. Optimization of
thermal masslevels dependson building material properties,
building orientation for its location and distribution, thermal
insulation, ventilation, climatic conditions and useof auxil-
iary cooling systems, and occupancy patterns. The most
important features,for eachof theseparameters,arepresented
in the following discussion.
2. I. Material thermal properties and petformance
The temperaturedistribution within the building materials
varies with time, boundary conditions and material thermal
properties.The phenomenawhich take placeduring day and
night periods, differ significantly, dependingon whether the
massmaterial is being charged (temperature increase) or
discharged(temperature decrease),respectively.
The thermophysical propertiesof heat storing material can
influence the performance of the system.For the material to
effectively store heat, it must exhibit a proper density, high
thermal capacity, and a high thermal conductivity value, so
that heat may penetrate through all the material during the
specific time of heat charging and discharging.
The depth that the diurnal heat wave reacheswithin the
storagematerial dependson thermal diffusivity (the control-
ling transport property for transient heat transfer). Materials
with higher thermal diffusivity values can be more effective
for cyclic heat storage at greater depth than materials with
lower values. Beyond a certain material thickness, the heat
flow into the indoor air doesnot take place during the night
hours, but it is delayed till the following daytime hours. This
is undesirableduring the cooling season,sincethe releaseof
storedheat to indoor spacesduring the early hoursof the day
will result in unpleasant indoor comfort conditions and
increaseearly day cooling loads.
2.2. Thermal masslocation and distribution
Location of thermal massis important. One may distin-
guish two cases,basedon whether the heat storagematerial
receivesenergy by solarradiation (direct) or by IR radiation
and room air convection (indirect). Direct heat gains are
experiencedby the outer building envelope, which is exposed
to solar radiation and by the interior surfaces which may
absorbincident solar radiation as it enters through the buil-
ding’s openings. Indirect heat gains are experienced by
opaqueelementsinside the building from the energy which
is transferred indoors from direct gain surfaces.Direct loca-
tions are much more effective than indirect, for placing heat
storagemass.
Thermal mass must be properly distributed around the
building dependingon the orientation of a given surfaceand
the desirabletime lag [ 51. North andeastsurfaceorientations
have little needfor time lag. For a south and west side,an 8
h time lag is sufficient to delay heat transfer from midday
until the evening hours. The roof, which is exposed to solar
radiation during most hours of the day, would have required
a very heavy construction with a long time lag. However,
becauseit is very expensive to construct massive roofs, the
useof additional insulation is usually recommendedinstead.
2.3. Thermal massand insulation
In general,both thermal massand insulation are important
in the overall thermal performance of a building [6]. In
climateswhere cooling is of primary concern, thermal mass
can reduceenergy consumptionprovided that the building is
unusedin the evening hours and the stored heat can be dis-
sipatedduring this idle period. However, common insulation
materialswill deteriorate the performance of a thermal stor-
age wall [7]. Overall, because thermal mass stores and
releasesheat, it interacts with the building operation more
than the simpleaddition of insulation [ 81. This makesana-
lyzing thermal massperformance and the overall building
thermal performancea more difficult problem to treat.
2.4. The role of ventilation
The role of thermal massis also extended into the night
period. During the summer,sinceoutdoor night temperatures
are usually lower than indoor temperatures,it is possibleto
cool the building by natural night ventilation. Ventilated air
enhancesconvective heat lossesfrom mass elementsand
dissipatesthe releasedheat to the lower temperatureoutdoor
 C.A. Balaras / Emrgy
heat sink. Ceiling fans can also be used to increase indoor air
movement and raise the convective heat transfer coefficient,
which facilitates the process of rejecting heat from massive
walls at night [ 91. ‘Ihe recommended night ventilation hourly
rate is 90 m3/h m2 of floor area. The objective is to maximize
the building thermal transmittance during the nighttime and
minimize it during daytime.
When the climatic conditions do not permit the use of
outdoor air, because of high humidity levels, one can still
precool the building during off-peak hours, using an air-
conditioning system. This, in fact, results in considerable
savings to the user and to the electric utility [ lO-131. The
electricity rates at night are usually much lower, since the
demand drops considerably. On the other hand, the utility
company also benefits, since eventually there will be a smaller
demand during the day hours, which in turn reduces the need
for additional generating capacity and maintains a more uni-
form load on the power generating facility. The saving is
achieved by subcooling the mass immediately before the peak
utility rate period, thereby shifting a portion of the cooling
load off-peak [ 121.
At the beginning of the following day, the cooled mass is
utilized as a heat sink. Part of the cooling load will be pas-
sively covered by the building mass which is at a lower
temperature, provided that the building is well insulated and
is not ventilated during the daytime. This means that for air-
conditioned buildings it is possible to reduce the energy con-
sumption of the system, since it will reduce its operation time.
For naturally ventilated buildings it is possible to achieve
longer periods within comfort conditions.
2.5. Occupancy patterns
A designer, should keep in mind that building occupants
constitute the final determining factor on the extent of util-
izability of any building system, including thermal mass.
Clearly, by changing the use of internal spacesand surfaces
one can drastically reduce the effectiveness of thermal stor-
age. Consequently it is necessaryto carefully consider the
final useof the spacewhen making calculationsof the cooling
load and incorporating the possible savings from thermal
masseffects.
3. Effectiveness of thermal mass
The effect of building thermal storageon the peakcooling
load has been evaluated during an experiment with a large
office building in Jacksonville, Florida [ 141. The objective
was to precool the building at night and during the weekend,
to reduce daytime cooling loads. Diurnal heat capacity cal-
culations were used in analyzing the experimental results.
The resultsshowedan 18% reduction in cooling energy sup-
plied during daytime.
A theoretical analysisof datacollected during an extensive
monitoring campaignof office buildings, using variable lev-
and Buildings 24 (1996) l-10 3
els of thermal massin similar buildings, hasprovided useful
information [ 151. Accordingly, an increaseof thermal mass
from 21 to 201 kg/m2 of floor area,in close and in ventilated
buildings, can reducethe peak indoor temperatureby approx-
imately 1 and 2 “C, respectively. Important reductionsin peak
heating and cooling loads can be achieved for high thermal
massdesigns.Energy savingsvary between 18 to 20%, com-
paredto a basecase.
A high masstest housewasmonitored in an environmental
chamberundera nighttime cooling operation [ 161. The cool-
ing load was measuredfor simple cyclic outdoor climatic
conditions. The simulated outdoor conditions were for a
deserttype climate (low temperaturesand high humidity at
night, and hot and dry conditions during the day). Nighttime
cooling was proved to be much more efficient than 24 h
cooling, resulting in over 50% decreaseof the total daily
electric power consumption. For cooling load calculations,
under these climatic conditions, it is however important to
account for a large amount of latent cooling load, resulting
from the release of moisture from the structural concrete
during the daytime warm-up period.
For indoor temperaturesset at 24 “C, heavyweight build-
ings have beenreported to consumelesscooling energy than
comparablelightweight buildings having the equivalent ther-
mal resistancein their walls [ 171. The thermal masswasalso
found more effective when it was positioned inside the wall
insulation. Nighttime ventilation can reducecooling loadsby
27 to 36%, dependingon the wall construction of the building.
Larger savingsoccur in heavyweight comparedto lightweight
buildings.
Robertson [41 haspresentedspecificrecommendationson
how to best use thermal massmaterials in residential and
small commercial buildings, for southwesternUnited States
locations. Denseand usually more conductive building mate-
rials are better. A 5-10 cm wall thicknessis usable for heat
absorption,storageandreleaseon a daily basis.Thicker walls,
upto 25 cm, may provide longer periods of thermal storage.
Thermostatsetupduring summer,is lesseffective in massive
structures,becausethey take longer to cool down. Numerical
simulationshave also shown that the percent annual reduc-
tions in sensibleheating and cooling loads, becauseof the
thermal masslevels of the walls (floors andfoundations were
assumedto be massless),can be as high as 40% in mild
climates.
For the designer,experimentalresultsprovide usefulinfor-
mation, but for actual building designone needsappropriate
tools to be able to deal with specific applications. There are
severaldesignand computational tools available, which take
into account the impact of thermal masson the building’s
thermal performance.In a recently completedstudy, 128pro-
grams have been identified, that have some capabilities to
calculate the cooling load of a building [ 181. Among them,
54 programscan alsocalculate the variation of the indoor air
temperature, 45 programs can also take into account the
impact of massand shading,while only 23 programscan also
simulatenatural ventilation strategies.
4. Simplified design tools
In total, 16 simplified models for estimating the cooling
load of a building, taking into account the building’s thermal
mass, have been selected and are briefly described in the
following discussion, in alphabetical order. Simplified design
tools are necessary for building designers during the Hurst
stages of a design process in order to evaluate new techniques
and systems and their performance under specific conditions.
These tools are easy to use, do not require detailed inputs and
provide relatively accurate information. They may require
the use of a computer or in some cases the calculations can
also be performed by hand.
The reviewed models are tabulated in Table I, which
includes some relevant information for each one. For each
method, one can identify the inputs and outputs, the parameter
used to account for the thermal mass effect, the restrictions
and whether the method has been used to develop a software
package. This review has also revealed a number of para-
meters which can be used to account for the thermal mass
effect on the cooling load and overall behaviour of buildings.
These parameters and their physical meaning are summarized
in Table 2.
4.1. Description of models
ASHRAE [ 191 has adapted a simplified design method.
The calculations can be performed by hand using a few fun-
damental equations but numerous tabulated input data. The
thermal mass effect on the cooling load is taken into account
by various factors, called cooling load temperature differ-
ences (CLTD) and cooling load factors (CLF) . These fac-
tors were generated for each component of the space cooling
load using the methodology and basic equations of the trans-
fer function method (TFM) Following the TFM procedure.
extensive computer simulations were first carried out, and
then based on these results, the representative multipliers
(CLTD and CLF values) were extracted by dividing com-
ponent results by their U value. The space cooling load com-
ponents are calculated directly, using tabulated CLTD and
CLF, which include the effect of time lag due to thermal
storage. The method can account for the significance of the
thermal mass on the cooling load for light, medium and heavy
building constructions, but it cannot directly handle the cal-
culation of optimum thermal mass level and distribution. An
additional limitation of the method is that the tabulated factors
have been calculated for a constant indoor air temperature,
an assumption which is not valid for buildings where tem-
perature swings are expected. A modification of the CLF
factor to account for external window shading, has been pro-
posed by Todorovic [ 201. This so-called { negative cooling
load’ method provides more accurate results with differences
that can be as much as 50% [ 2 I].
The Athienitis model [ 22,231 is particularly suitable for
massive buildings. It is a predictive control algorithm, which
utilizes computer prediction of the building’s response to
expected weather inputs to modify its operation strategy so
as to reduce energy consumption and temperature swings
resulting from large solar and other radiant gains, and highly
variable loads. The model is developed for a heating season.
The building response to expected weather inputs and internal
gains is predicted using a frequency domain thermal network
model. Using an iterative technique, a profile of the thermo-
stat set point is determined so that optimum comfort and
minimum energy consumption are achieved. The building
model employs distributed elements, such as thermal storage
mass in walls, accounts for thermal coupling between rooms,
room interior radiant exchanges, and a time varying
conductance.
The Bida and Kreider model [ 241 is a simplified design
method which can be used for calculating the monthly cooling
loads of dwellings and small multistorey buildings, with small
internal gains. The main source of heat which is transferred
into the building is from incident solar radiation. The interior
of the building is considered as one zone. The component of
the cooling load from the latent heat is not taken into account.
However, the model accounts for loads from ambient and
internal gains, and solar gains through transparentand opaque
elements. The mass effect coefficient (MEC) is used to
account for thermal mass effects on the cooling load and is
defined as the ratio of the calculated monthly cooling load to
the monthly cooling load if no thermal mass were present.
The model can estimate the annual cooling loads with an
acceptable accuracy. For large cooling loads, greater than 10
GJ/year, the maximum error is less than 15%. The calcula-
tions of the cooling load can be performed for light and heavy
buildings, thus enabling the user to easily compare the effect-
iveness of the thermal mass in reducing the cooling load of a
building.
The Building Research Establishment [ 251 has developed
a method for calculating peak summer temperatures. The
method introduces the concept of the admittance factor. This
represents the extent to which heat enters the surface of mate-
rials in a 24 h cycle of temperature variation. This follows
from the fact that dense materials take up more heat and have
higher admittance values than lightweight ones. The method
has been developed in an easy to follow format of a series of
worksheets that the user has to fill out and includes the effect
of material thermal capacity and resistance. It is suitable for
office buildings, with one external wall and all other faces
about other offices, but it can be extended for treating rooms
not surrounded by others, with a reduced accuracy.
Catani [ 261 has developed a computer program account-
ing for thermal mass effects. The method is known as the M
factor, which is a correction to steady-state R values that can
result in less insulation being required in opaque wall sections
of massive buildings than for lightweight buildings. Charts
of M factors show the correction as a function of local degree
days and the weight per unit of wall area. The h4 factor is
used to indicate the thermal performance of building enve-
lopes and as a correction to the steady state U values, which
are divided by this factor. Accordingly, the M factor is a
 C.A. Balarus /Energy und Buildings 24 (1996) l-10

Table I
Simplified design tools for calculating the cooling load in buildings

Model Inputs Outputs Parameter to describe Restrictions Software

mass thermal effects

1. ASHRAE [ 191 BG,BC,IL,EL,AT/M,VE/S and L, hourly and peak cooling CLTD values for wall simplified treatment no
loads construction
Todorovic [ 211 IN/S and L modified CLF for
windows
2. Athienitis [22] BG,BC.AT/M.ATIH.SRIM or hourly or monthly for elements in a thermal detailed inputs, yes
network complex analysis
SR/H,IL optimum thermostat set
point
3. Bida and Kreider BG,BC,IL,EL.VEIS,SRIM. monthly loads mass effect coefficient monozone, dwellings. yes
[241 small multistorey
AT/M,CL,SS buildings, small IL, no
latent heat loads, no
wind effects
4. BRE [25] BG,BC.EL,IL peak summer admittance factor office buildings with no
temperatures one external wall. no
wind effects
5. Catani [ 261 BG,BC performance of masSlYe M factor manual method for no
walls calculating the
thermal mass
performance
6. Givoni [27] and BG.BC long term performance of effective heat capacity Method for calculating yes
Balcomb [ 71 thermal mass
massive walls performance
7. Givoni and BG,BC,VE,SRIM,AT/M internal air temperature total thermal time All external input yes
Hoffman [ 301 constant functions are taken
equal, no wind effects
8. Kusuda [31] BG,BC,EL,IL,VE,TAIH,SRiH transient sensible cooling conduction transfer Detailed inputs, no
load functions complex analysis for
previous temperature history multi-zone,-element
cases
9. LANL [ 361 BG,BC,SR/M,DD monthly cooling load, solar saving fraction Restricted to certain yes
daily type of buildings, no
average indoor VIZ IL, wind effects
temperature
10. van der Maas BG,BC.IL,AT,VE/S indoor air temperature, thermal effusivity Monozone. stack effect yes
and Roulet [ 371 driven flows, short
cooling rate periods, homogeneous
walls, no radiation,
no convective effects,
constant heat flow
rates, no wind effects
I 1. Mathews and BG,BC,IL,EL.VE,SR/M. hourly Indoor an effective heat storage Single zone, no yes
Richard [ 381 temperature temperature
stratification,
AT/M and sensible loads interior surfaces at
same temperature, no
wind effects
12. Passport Plus BG,BC,EL.IL,IN/S/L,VE/S hourly indoor air and transfer functions One dimensional heat yes
1411 surface conduction
ATIH,SRIH,WVIH,WDIH temperatures, cooling/
heating
(Meteorological Library files) load, air flows, comfort
13. PMDG [36] BG,BC,SR,DD monthly cooling load, heat capacity no VE, IL. wind yes
daily effects
average cooling load
14. Seem [33] Previous temperature and heat cooling load comprehensive transfer Calculations for input no
functions information
flux, transfer functions, loads for
AT.IL,EL

( conlinued)
6 C.A. Balarus /Energ.v und Buildings 24 (1996) I-10

Table I
(continued)

Model Inputs outputs Parameter to describe Restrictions Software

mass thermal effects

15. Shaviv [43,44] BG,BC,SR/H.AT/H maximum indoor air heat capacity, thermal isothermal internal no
resistance mass
temperature
16. SPIEL [46] BG,BC,SS,AT/M,IN mean moonthly hourly capacitance convergence of
indoor calculated values, no
air temperature, cooling direct account of SR
load levels, no IN, no
wind effects
17. Steady State BG,BC,AT/M.lN average daily space envelope heat transfer steady state conditions, no
[I71 cooling coefficient no IL,EL,VE

Nomenclature: ADT = monthly ambtent daytime temperature; AT/M, AT/H = ambient temperature monthly data, ambient temperature hourly data;
BC = detailed information on the building construction ( U values, material thermal properties) ; BG = detailed information on the building geometry, orientation;
CDD = monthly cooling degree daysCL = monthly clearness index; DD = monthly degree days; EL = external loads (solar radiation, temperature); IL = internal
loads (human activity, equipment, lights); IN/S, IN/L = infiltration loads from sensible heat, infiltration loads from latent heat; SR/M, SR/H =solarradiation
monthly data, solar radiation hourly dataSS = mean daily sunshine hours; VE/S, VE/L = ventilation loads from sensible heat, ventilation loads from latent
heat; WV/M, WV/H= wind velocity monthly data, wind velocity hourly data; WD/M, WD/H= wind direction monthly data, wind direction hourly data.

Table 2
Parameters for describing thermal mass effects

Parameter Phystcal meaning Ref.

Admittance factor represents the extent to which heat enters the surface of matenals in a 24 h cycle of 1251
temperature variation
Capacitance accounts for the ability of the external and internal materials to store heat r491
Comprehensive transfer functions describes heat flows in building elements, combining individual wall transfer functions [331
for an enclosure
Conduction transfer functions expresses the decay of temperature through the material [311
Cooling load temperature difference includes the effect of time lag in the propagation of heat through the material, due to [I91
thermal storage
Diurnal heat capacity measures the effectiveness of the material for heat storage during a continuous 24 h [281
cycle
Effective heat capacity accounts for the effects of the building’s materials’ thermal properties and design [291
factors on the long term energy performance
Effective heat storage accounts for the effects of thermal capacity and thermal resistance of the building [381
elements, and the extenorresistance. for exterior and interior building elements
Envelope heat trasnfer coefficient includes the effects of thermal transmittance of the material along with heat transfer rate Cl71
due to infiltration
Heat capacity introduces the effect of heat storage for different building types in the correlation [361
coefficients used in calculating the solar saving fraction
Mass effect coefficient accounts for the temperature fluctuation allowable in the space, the amount of heat gain v41
due to ambient conditions, and the degree of exposure of the space to ambient
conditions
M Factor corrects steady state U values for the building materials P61
Solar saving fraction correlation coefficients as a function of the heat capacity for specific building types [361
Thermal capacity determines the heat flow in unit time by conduction through unit thickness of a unit area
material. across aunit temperature gradient, defined as the product of density by spesific
heat
Thermal effusivity accounts for the response of a surface temperature to a change of the heat flow density [371
at the surface
Total thermal time constant the heat stored m a whole enclosure per unit of heat transmitted to or from the outside [301
through the elementssurrounding the enclosure and by ventilation

dimensionless correction factor, not a direct measure of ther-
ma1 storage or insulation capacity of walls.
The model of Givoni [ 271 and Balcomb and Jones [ 71
utilizes the concept of heat capacity, the product of density
by specific heat, of building materials. The model distin-
guishes between direct and indirect gains in thermal mass, by
using the diurnal heat capacity (dhc) and the total heat capac-
ity (the) . The dhc, introduced by Balcomb and Douglas [ 281,
is a measure of material heat storage effectiveness during a
continuous 24 h cycle, while the measures the total heat stored
 C.A. 6alarct.s / Energy
if all the material was uniformly raised in temperature. In the
Givoni-Balcomb model a similar parameter called the effec-
tive heat capacity (ehc) is introduced to predict the effects
of material properties (density, thermal conductivity and spe-
cific heat) and design factors such as the surface area, thick-
ness and location of the structural storage elements, on the
long term energy performance of a passive solar system. The
model has been validated with results obtained from other
models and can predict satisfactorily the performance of
direct gain buildings.
The effective heat capacity and a generalized solar load
ratio correlation have also been used in a method developed
by Wray and Best [ 291. The effective heat capacity is first
defined empirically, by calculating the quantity that best cor-
relates with auxiliary energy consumption, and then new solar
load ratio correlations are established, that contain the result-
ing effective heat capacity as the single variable pertaining
to thermal storage. This work confirms the hypothesis by
Givoni that ehc resembles dhc, but has no optimum relative
to thickness, for a fixed area.
Givoni and Hoffman [ 301 extended the use of the thermal
time constant (TIC) concept for an enclosure, by introduc-
ing the total thermal time constant of the building (TITCB),
for measuring the ability of the enclosure’s interior mass to
admit and store heat. Originally, TTC was derived from the
analogy between the heat flow through the building’s mate-
rials represented by a thermal circuit and the time constant of
an electric circuit, as a single parameter that can be used for
evaluating the thermal performance of a building. The
TTTCB is defined as the heat stored in the whole enclosure,
including internal air, per unit of heat transmitted to or from
the outside through the elements surrounding the enclosure
and by ventilation. The shorter the TTTCB, the smaller the
heat flux needed to achieve a unit variation of temperature.
The method calculates internal air temperature and has been
validated for different model building materials and geome-
tries, under natural climatic conditions and with full-scale
buildings in different climatic zones [ 301.
Kusuda [ 3 1] has developed a model based on the response
factor method (RIM) [ 321. It utilizes the conduction trans-
fer functions (CTF), which is a convenient and effective tool
for the evaluation of transient heat transfer through building
construction components. The RFM and CTF methods are
used in building simulation programs to solve transient heat
transfer problems in which the heat transfer properties are
time-invariant. The advantage of these methods is that they
are more efficient than other classical techniques for solving
long-term heat transfer problems like Euler or Crank-Nicol-
son, because there is no critical time step and the internal
temperature distribution is not calculated [ 331. The RFM has
also been used by Stephenson and Mitalas [34] for one-
dimensional problems. An exact set of transfer functions and/
or response factors is determined by solving the conduction
equation by Laplace and/or z-transforms. Ceylan and Myers
[ 351 have presented a method for calculating transfer func-
tions for multi-dimensional heat transfer from a set of first-
and Buildings 24 (I 996) I-10 I
order differential equations, which requires the calculation of
eigenvalues and eigenvectors of a matrix. Seem et al. [ 331
have improved the last method, by eliminating the interme-
diate steps, thus providing a simple computational and accu-
rate two- or three-dimensional heat transfer calculation. The
basic heat transfer process that occurs in a building zone is
illustrated by an electrical circuit network. The heat transfer
system is then solved in the zone temperature or the zone
load calculation modes. The Kusuda model has been used for
a study of various types of buildings. The cooling require-
ments of the buildings studied in this series averaged 22%
less for heavier constructions.
Researchers at the Los Alamos National Laboratory
(LANL) have published numerous data on tests they have
conducted on passive test cells, along with various correla-
tions which were developed using a mainframe computer
program (Passive Solar Energy code, PASOLE). The use of
the LANL model, however, is limited due to the fact that it
is applicable only to buildings similar in their heat capacity,
type of glazing or night insulation, to the types specified by
the model. The model utilizes the solar saving fraction [ 361
which depends on the building’s thermal storage capacity,
nominal mass thickness, mass to glazing area ratio, number
of glazing and night insulation conditions.
van der Maas and Roulet [37] have developed a model
for simple energy calculations of stack ventilative night cool-
ing in a building zone, with an air-entrance and air-exit open-
ing. The algorithm links stack ventilation to heat transfer and
a wall thermal model and predicts the influence of building
materials and opening sizes on the heat transferred out of one
zone by nighttime ventilative cooling. The heat transfer proc-
esses are represented by a thermal network. Themethod intro-
duces the thermal effusivity of building materials (the square
root of the product of density, specific heat and thermal con-
ductivity) which accounts for the response of the surface
temperature to a change of the heat flow density at the surface.
The model has been validated with measurements collected
during tests conducted in office rooms ventilated by a single
open window, and a test at a high mass, three level, staircase
in an office building, cooled by stack ventilation.
Mathews and Richards [ 38,391 have developed a model
for thermal simulation of a building zone. It predicts hourly
air temperatures and sensible energy loads and treats the
effects of heat storage. The proposed approach introduces a
novel treatment of heat storage in buildings and a term
describing varying ventilation rates. The relevant equations
are solved under the assumption that all variables are periodic.
The thermal network for a single building zone includes the
time-dependent radiative heat source acting on the internal
structure, the solar temperature acting on the exterior, the
convective heat source acting on the internal air volume, and
the outdoor temperature acting on the internal air volume.
The method uses a single zone approach and consequently
no heat exchange exists between different zones in a building.
The air temperature distribution throughout the zone is
assumed uniform. All interior surfaces are assumed to be at
8 C.A Balurus /Energy
the same temperature. The model has been extensively veri-
fied in more than 60 passive, naturally ventilated buildings
[40]. For 80% of the time temperature predictions were
within 2 “C of measurements, showing that heat storage
effects are efficiently simulated.
Passport Plus [ 411 is a simplified computational tool, for
DOS compatible micro-computers, developed within the
CEC Research Programme on passive cooling in buildings,
PASCOOL [ 421. The program is structured in such a way as
to easily accept enhancements and modifications, in the form
of a routine that treats specific phenomena. Some of the pro-
gram’s features include a more detailed treatment of the ther-
mal mass problem based on transfer functions and finite
differences, accounting for the role of the building’s sur-
rounding external environment, small scale micro climate,
external remote obstacles, external shading devices like lou-
vres, and improved treatment of natural ventilation phenom-
ena. The program handles multi-zone buildings, including
active and inactive zones, using hourly input meteorological
files. Passport Plus calculates indoor air and wall tempera-
tures, indoor air flows, cooling or heating loads and indoor
thermal comfort conditions. The program has undergone an
intermodel comparison between several well-known models
and with experimental data from several European sites, with
satisfactory results.
Givoni [36] has developed a new predictive model for
direct gain (PMDG), to replace the constants of the LANL
model with functions dependent upon the thermal properties
of the building, the components of the solar system, or both.
The modified LANL model can be used for buildings with
any heat capacity, down to a lower limit of 150 Wh/m’ “C.
Based on an experimental study and a comparison between
the PMDG and LANL models with the collected data, it
appears that the LANL mode1 can be used even for buildings
that do not have the exact properties of the reference build-
ings, although the PMDG model may provide a more accurate
prediction of the auxiliary load.
Seem et al. [ 331 have developed a model which utilizes
the TFM procedures for computing the cooling loads and
floating room temperatures in buildings. The transfer func-
tions describing heat flows in building elements are combined
in a comprehensive room transfer function (CRTF) for an
enclosure. CRTF simulations require less computational
effort than energy balance simulations because only outputs
of interest are computed. The method accurately models long-
wave radiation exchange and convection in an enclosure
through an approximate network, called the { star’ network.
Resistances in the star network are determined from a network
which uses view factors to mode1 long-wave radiation
exchange. Combining heat flows from every surface to the
star nodes results in the CRTF equation, which relates the
load for an enclosure to past and current inputs and past loads.
The Shaviv model [43,44] is a time-dependentmethod
which simultaneously solves the heat transfer equation
through all exterior walls of the building, taking into account
the thermal massof exterior walls andpartitions. The internal
und Buildings 24 (1996) I-IO
thermal mass is treated as an isothermal mass. The mode1
calculates the maximum indoor air temperature, without any
mechanical means, or with a fan only. A series of simulations
using this model has shown that the optimum thickness of a
concrete layer of an internal partition should be 10 cm, while
the recommended thickness of an external wall is 15 cm [45].
SPIEL [ 461 is a simplified design tool for buildings which
can calculate the mean monthly hourly indoor temperature
and required backup heating or cooling to reach the set indoor
air temperature. This approach does not allow for adistinction
between direct or indirect thermal gain mass, thus it is not
possible to optimize the location of the mass. Thermal mass
effects are taken into account by the capacitance of the build-
ing, that is, its ability to store heat. The mode1 can handle
multi-zone buildings, it accounts for internal heat gains,
natural and forced ventilation, and indirectly for night ven-
tilation. A new version of the program, recently released, has
improved its many features.
Finally, according to the simple steady-state model [ 171,
when buildings are exposed to hot summer weather condi-
tions for which both the day and night outdoor temperatures
are sufficiently warm to cause a 24 h continuous net positive
space cooling load to exist, the average daily space cooling
load can be calculated by the steady-state theory. This relation
is the product of the envelope heat transfer coefficient and
the temperature difference of the average outdoor tempera-
ture and a balance point temperature which is the average
daily outdoor temperature at which the daily space cooling
load vanishes.
Simplified design tools offer an alternative to more
advanced and complex programs which are also available.
Advanced models require a great amount of input information
and provide in return the capability for treating complicated
phenomena with high accuracy. Advanced programs are suit-
able for simulating the behaviour of buildings and evaluating
the performance of various systems with great accuracy. They
are more likely to be used by people with advanced knowl-
edge on these subjects, for scientific and research activities.
There are several advanced computer models for simulat-
ing the thermal behaviour of buildings, like DOE [ 471, TARP
[48] and BLAST [49]. Several European research pro-
grammes in the area of energy conservation in buildings, have
utilized extensively the environmental systems performance
(ESP) software [ 505 11. TRNSYS [ 521 has also been a very
popular commercially available program in the field of solar
energy system design, mainly because of its modularity, espe-
cially following the release of a version for persona1
computers.
5. Conclusions
The massstructure plays an essentialpart in the thermal
responseof buildings. Generally, a high massbuilding hasa
smaller interior air temperature variation than a low mass
building, while there is a decreaseand a time lag of the peak
 C.A Bolaras / Emrgy und Buildings 24 (19961 l-10 9

cooling load. For locations with large diurnal temperature [13] M. Brandemuehl, M. Lepore and J. Kreider. ASHRAE Trans., 96
(1990) 871-877.
swings, this process can significantly reduce the energy con-
[ 141 M. Ruud, J. Mitchell and S. Klein, ASHRAE Trans.. 96 ( 1990) 820-
sumption of the mechanical cooling system. This technique
835.
is more applicable to offices and other buildings which are [ IS] M. Brown, S&r Energ)) Eng. 112 ( 1990) 273-282.
unoccupied during the night, so that the structure can be [ 161 T. Kusudaand J. Bean,ASHRAE~rclns.. R7 (1981) 1232-1240.
cooled with nighttime ventilation. Air-conditioiled buildings [ 171 D. Burch, S. Malcolm and K Davis. ASHRAE Trans., 90 ( 1984) S-
can also be precooled during off-peak hours, for considerable 21.
energy savings. Resulting small indoor temperature varia- [IS] M. Antinucci. B. Fleury. J. Lopez d’Asiain, E. Maldonado. M.
Santamouris, A. Tombazis and S. Yannas, Int. J. Solar Energy, I1
tions have also a positive influence on occupant thermal
( 1992) 251-262.
comfort. [ 191 ASHRAE Handbook of Fundamenruls, American Society of Heating
For the designer, the primary objective is to calculate the Refrigerating and Air conditioning Engineers, Atlanta, GA, 1981.
optimum thermal mass of a building and to then distribute it [ 201 B. Todorovic, ASHRAE Trans., 88 ( 1982) 924-937.
in such a way as to optimize temperature fluctuations and [2l] B. Todorovic, ASHRAE Trans., 99 (1993) 41-46.
heat return rate into the indoor environment. Material selec- [22] A. Athienitis, ASHRAE Truns., 94 ( 1988) 1050-1067.
[23] A. Athienitis. M. Stylianou and J. Shou, ASHRAE Trms.. 96 (1990)
tion is also important, since thermophysical properties influ-
839-848.
ence the overall performance of heat storage systems. [24] M. Bida and J. Kreider. Solar Energy Eq., 109 ( 1987) 31 I-320.
The evolution of cooling load and energy calculations from [ 25 I P. Burberry, Prrrcr~ul Thermcll Desqn in Buildings. Batsford
the crude whole building approximations 30 years ago to the Academic and Educational Ltd., London, 1983.
current sophisticated analysis offered by many commercially [ 261 M. Catani, ASHRAE J., 21 (June) ( 1978) 50-55.
available programs has been overwhelming [ 531, Amongst [ 271 B. Givoni, Passrve Solar J.. 4 ( 1987) 25-110.
[28] J. Balcomb and J. Douglas, Corlcrere Musonry Solar Archif. Q., 2
the several available design tools, for cooling load calcula-
(1982) 67-89.
tions, the user has to select one based on specific criteria for [29] W. Wray and E. Best, Passive SolurJ.. 4 (1987) 41-61.
a given application. Depending on the features of a model, [ 301 B. Givoni, Man, Clrmate and Architecture. Applied Science, London,
available information and desired accuracy, one can identify 2nd. edn., 1981.
the most appropriate tool for practical applications or a more [3 I] T. Kusuda, Heat transfer in buildings, in W. Rohsenow. J. Hartnett and
in-depth analysis. E. Ganic (eds.). Handbook of Hear Transfer Applicafions, McGraw
Hill, New York, 1985. Ch. 9.
321 T Kusuda. ASHRAETruns.. 75 ( 1969) 250-269.
331 J Seem. S Klein, W Beckman and J. Mitchell. Comprehensive room
Acknowledgements transfer functions for efficient calculation of the transient heat transfer
processes in buildings, ASME/AICHE National Heat Transfer Conf,
Pirrsburgh. PA, Aug. 1987. Technical Session on Heat Transfer in
This work was conducted in the frame of the PASCOOL, Buildings, pp. 25-33.
CEC Research Project, which was partly financed by the 341 D. Stephenson and G. MitaIas, ASHRAE Trans.. 73 (1967) 3&41.
European Commission.It is publishedwith permissionfrom 1351 H. Ceylan and G. Myers, ASME J. Heor Transfer, 102 ( 1980) l15-
the PASCOOL coordinator, Dr M. Santamouris,University 120
[ 361 B. Givoni, Passive Solar J.. 4 ( 1987) S-24.
of Athens.
[ 371 J. van der Maas and C. Roulet, Night time ventilation by stack effect,
in Ventilation andInfiltrotwn, ASHRAE Tech. Data Bull., Vol. 7 ( I ),
NY-9 I -S-3, ASHRAE, Atlanta, GA, 1991. pp. 3240.
References [38] E. Mathews and P. Richards, Energ)] Build., 14 (1990) 61-72.
[39] E. Mathews, P. Rousseau, P. Richards and C. Lombard, Build.
Environ.. 26 ( 1991 ) 179-186.
[ 11 M. Brown, Solar Enei-,q?, Eng., 112 ( 1990) 273-279. [ 401 Quick, a thermal analysis program, Version 4. Centre for Experimental
[2] R. Kammerud, E. Ceballos, B. Curtis, W. Place and B. Andersson, and Numerical Thermoflow, Department of Mechanical Engineering,
ASHRAE Trans.. 90 ( 1984) 226-252. University of Pretoria, Republic of South Africa, 199 1.
[3] S. Szokolay. Passive and low energy design for thermal and visual
[4I] C.A. Balaras and S. Alvarez, Pussport Pius. User’s Manual, Pascool
comfort, Passive and Low Energy Ecorechniyum 3rdlnt. PLEA Corlf., Project, Joule II Programme, JOU2-CT92-0013. European
Mexico Ciry. 6-11 Aug. 1984, pp. 11-28.
Commission, 1995.
[4] D. Robertson, Pussive Solar J., 3 ( 1986) 387417.
[42] M. Santamouris and A. Argiriou, The CEC Project Pascool, 3rd
[5] N. Lechner, Heating, Cooling. Lighting, Wiley, New York, 1991.
European Conf: Architecture. Florence. Italy, 17-21 May. 1993, pp.
[6] V. Hopkins, G. Gross and D. Ellifritt, ASHRAE Trans.. R5 ( 1979)
17-21
885-902.
[431 E. Shuviv and G. Shavlv, Comput. AldedDes. J.. 10 (1978) 239-247.
[7] J. Balcomb and R. Jones, Workbook f;jr Workshop m Adwumd
Passive Solur Design. Bled, Yqo.&via, Balcomb Solar Associates. [44] E. Shaviv and G. Shaviv, Build. Environ.. 13 ( 1978) X-108.
1988. [4S] E Shavlv. On the determmation of the optimum thermal mass in the
[ 81 D. Ober and D. Wortman, ASHRAE J.. 33 ( 1991) 18-2 I medlterranean &mate. in E. de Ohveira Fernandes and S. Yannas
[9] S. Baer. Raising the open U value by passive means. Proc. 8th Nuiclr (eds.), Energy and Buildingsfhr Temperare Climates, 6th Int. PLEA
Passive Solar Conj, Glorieta. NM, 1983. pp. 839-842. Co@. form. Portugal, Pergamon, New York, 1988, pp. 385-390.
[IO] J. Braun, ASHRAE Truns., 96 ( 1990) 876-888. [461 C. Green, The simulation tool SPIEL. Ecotech, UK.
[II] C. MacCracken,ASHRAEJ.,33 (Dec.) (1991) 12-23. [47] S. Diamond, DOE Program Manual, Rep. u-7688. Los Alamos
[ 121 M. Snyder and T. Newell, ASHRAE Tram.. 96 ( 1990) 830-836. Scientific Laboratory. Feb. 1979.
10 C.A. Balaras / Energy and Buildings 24 (I 996) I-IO

[48 J G. Walton, Thermal Analysis Research Program, Reference Manual, [SO] J. Clarke, Energy Simularion in Building Design, Adam Hilger, Bristol,
NBSIR 83-2655, National Bureau of Standards, Gaithersburg, MD, UK, 1985.
1983. [ 5 1 ] J. Clarke, Solar Wind Tech.. 6 ( 1989) 345-355.
[49J D. Hittle, BLAST, The Building Load Analysis and Systems [52] S. Klein, TRANSYS, a transient system simulation program, Rep. JR-
Thermodynamics Program, Vol. 1, User’s Manual, Rep. CERL- 12. Engineering Experiment Station, University of Wisconsin-
TR-E-JS3. US Army Construction Engineers Research Laboratory, Madison, 1988.
1979. [53] J.M. Ayers and E. Stamper, ASHRAE J., (Feb.) ( 1995) 47-55.